Underwear Pattern Matching and Typesetting Method Based on Machine Vision

Abstract

The present invention discloses a method for flower alignment and layout of underwear based on machine vision. The present invention relates to the technical field of underwear processing, and solves the problem that the original layout processing method cannot perform flower alignment and layout processing on underwear of different sizes. The present invention selects the starting layout line based on the contour features of the flower alignment layout area, and then matches and combines it with the printing starting line of the flower alignment template; by proportionally enlarging the flower alignment template and adjusting the angular features to make the two coincide, the complete flower alignment area is finally determined; this process ensures that the flower alignment template can be accurately applied to the flower alignment layout area, realizes efficient and accurate flower alignment and layout processing, improves the quality and efficiency of underwear flower alignment and layout, and can meet the requirements for fine flower alignment and layout in the production process.

Description

Technical Field

[0001] The present invention relates to the technical field of underwear processing, and specifically to an underwear pattern alignment and layout method based on machine vision. Background Art

[0002] When producing underwear, in order to keep the patterns, patterns, etc. on the surface of the underwear coherent, symmetrical and coordinated after splicing and sewing, it is necessary to accurately arrange and position the patterns on each cut piece. This process is the underwear pattern alignment and layout; for example, for a bra with a floral pattern, after sewing different cut pieces such as the cup surface and the side wings, the floral pattern should be naturally connected, looking like a complete painting rather than being disordered and incomplete.

[0003] The application with the patent publication number CN113408237B discloses a fancy layout method for improving the cutting utilization rate. The fancy layout shape is similar to a square shape and is symmetrical. Then, combined with the size of the SET dimension, it is placed horizontally or vertically and matched according to the size of the large material; the present invention provides a fancy layout method for improving the cutting utilization rate, which improves the utilization rate of the board and reduces the cost of the board through the fancy layout method.

[0004] During the process of pattern alignment and layout processing of its underwear, the associated area to be laid out is laid out based on the corresponding preset template. However, each different area to be laid out with different sizes needs to correspond to a preset template with different sizes. During the comparison processing, this method is relatively cumbersome and has low efficiency. At the same time, it cannot perform pattern alignment and layout processing on underwear of different sizes, and its original pattern alignment and layout processing method is not comprehensive enough. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the present invention provides an underwear pattern alignment and layout method based on machine vision, which solves the problem that the original layout processing method cannot perform pattern alignment and layout processing on underwear of different sizes.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: An underwear pattern alignment and layout method based on machine vision, including the following steps:

[0007] Step 1: Place the underwear to be pattern-aligned and laid out at a designated position, use a high-definition probe to obtain the surface image of the underwear, and perform preliminary processing on the obtained surface image, and select the complete image belonging to the underwear. The specific sub-steps are as follows:

[0008] S11: Convert the surface image of the underwear obtained by the high-definition probe, convert the original color image into a grayscale image, confirm the gray values associated with different gray points in the converted grayscale image, and label the different gray values as H i , where i represents different gray points;

[0009] S12. Check the gray value H of different gray points i against the preset value Y1, and mark the gray points with H i ≤Y1 as background points; otherwise, no marking is performed;

[0010] Remove the background area associated with the background points from the underwear surface image, and record the image after the removal process as the complete underwear image;

[0011] Step 2. Re-convert the determined complete underwear image into a color image, then perform feature contour processing on the converted complete underwear image, identify the pixel values associated with different pixel points in the complete underwear image, and based on the gradient performance of different pixel points, confirm the feature contours existing in the corresponding complete underwear image and confirm the contour areas. The specific sub-steps are as follows:

[0012] S21. Mark the pixel values associated with different pixel points in the corresponding complete underwear image after the conversion process as X k , where k represents different pixel points. Then, based on the adjacent pixel points existing around the corresponding pixel points, confirm the pixel values associated with the adjacent pixel points, and use the Sobel algorithm to process the confirmed groups of pixel values to confirm the horizontal gradient HT k and the vertical gradient ZT k associated with the corresponding pixel points;

[0013] S22. Based on the confirmed horizontal gradient HT k and the vertical gradient ZT k , lock the combined gradient associated with the corresponding pixel points, and use to confirm its combined gradient ZH k . Compare the confirmed combined gradient ZH k of the corresponding pixel points with the preset value Y2. When ZH k >Y2, mark the corresponding pixel points as gradient pixel points; when ZH k ≤Y2, no processing is performed;

[0014] S23. Based on the confirmed groups of gradient pixel points in the complete underwear image, connect the adjacent gradient pixel points, sequentially confirm the feature contours existing in the complete underwear image, and then mark the specific areas included in each different feature contour as contour areas;

[0015] Step 3. From the confirmed groups of contour areas in the complete underwear image, first determine the center points associated with the corresponding contour areas, then align the contour areas with the preset feature areas and perform coincidence degree analysis simultaneously, and select the pattern alignment area in the complete underwear image. The specific method is as follows:

[0016] S31. Based on the different edge contours associated with different contour regions, place the corresponding edge contours in a two-dimensional coordinate system, confirm the two-dimensional coordinates associated with the corresponding contour points, perform a mean process on the several groups of two-dimensional coordinates of the several groups of contour points associated with the corresponding edge contours, confirm the mean coordinates, and mark the points associated with the determined mean coordinates in the corresponding contour region, denoted as the center point of this contour region;

[0017] S32. Sequentially confirm the center points inside different contour regions, align the contour regions after the center point confirmation with the preset feature region, where the preset feature region is the originally planned typesetting region, and after alignment, rotate the contour region according to the center point or the feature region according to the center point. There is also a preset center point inside the feature region, and during the rotation process, record in real time the overlapping region generated between the contour region and the feature region, confirm the area ratio of the overlapping region located in the feature region, and use: ZB p = area of the overlapping region ÷ total area of the feature region, where p represents different rotation processes. When the contour region or the feature region rotates 360° according to the corresponding center point, stop rotating;

[0018] Based on the different area ratios ZB associated with different rotation processes p , select ZB p max as the characteristic value associated with this contour region;

[0019] S33. Sequentially confirm the characteristic values ZB p max of different contour regions, and mark the contour regions that satisfy: ZB p max ≥ 95% as the pattern alignment typesetting area. Otherwise, do not perform any marking;

[0020] If there is no contour region in the complete underwear image that satisfies ZB p max ≥ 95%, directly generate an error signal for display;

[0021] Step Four. Based on the determined pattern alignment typesetting area and the preset pattern template, first determine the relevant characteristics of the overall contour of the pattern alignment typesetting area, select the starting typesetting line from the pattern alignment typesetting area, and then based on the preset printing starting line in the pattern template, combine the pattern template with the pattern alignment typesetting area, and display the combined complete pattern area. The specific method is as follows:

[0022] S41. Confirm the contour of the determined pattern alignment typesetting area, lock the edge contour belonging to this pattern alignment typesetting area, then confirm the point angle between the corresponding contour points from the locked edge contour, and mark the confirmed different point angles as Jo, where o represents different corresponding contour points;

[0023] The corresponding contour points satisfying Jo≤150° are calibrated as feature inflection points, otherwise, no calibration is performed;

[0024] Confirm the length of the partial edge contour associated between adjacent feature inflection points, and calibrate the confirmed length as CL q , where q represents different partial edge contours, from the multiple groups of lengths CL q selected, select the partial edge contour associated with CL q max, and denote it as the starting layout line of this pair of pattern matching area;

[0025] S42. Based on the preset pattern matching template, select the starting line of printing from within this pattern matching template, calibrate the length of the starting line of printing as YL, and calibrate the length of the starting layout line within the pattern matching area as QL;

[0026] If YL < QL, then scale up the pattern matching template proportionally until YL≥QL for the length of the starting line of printing;

[0027] If YL≥QL, then no processing is performed;

[0028] S43. Confirm the center point Z1 of the starting line of printing and the center point Z of the starting layout line, make Z1 coincide with Z2, and based on the angular features between the center point Z2 within the starting layout line and adjacent points, adjust the angular features between the center point Z1 of the starting line of printing and adjacent points. Subsequently, using the center point as a reference, sequentially adjust the angular features of the surrounding points, and perform adjustments step by step in this way until a set of curves within the starting line of printing completely coincide with the starting layout line or the starting line of printing completely coincides with the starting layout line, and then stop;

[0029] After the starting line of printing completely coincides with the starting layout line, make the pattern matching template coincide with the pattern matching area, select the coincident area, and use the selected coincident area as the complete pattern matching area and display it.

[0030] The present invention provides an underwear pattern matching layout method based on machine vision. Compared with the prior art, it has the following beneficial effects:

[0031] By using the Sobel algorithm to confirm the contour area based on pixel point gradients, the present invention can accurately identify the inner and outer contours of underwear. This analysis method based on pixel point features fully considers the variation law of pixel values in the image and has good adaptability to the contour features of different types of underwear, so as to accurately divide the underwear area and provide a reliable basis for determining the pattern matching area;

[0032] By calculating the center point of the contour area and aligning and analyzing the overlap with the preset feature area, the pattern layout area can be scientifically selected. This method can effectively select the appropriate layout area based on the actual contour characteristics of the underwear and the preset layout requirements; if the area that meets the requirements is not found, an error signal is issued, which helps to promptly discover problems such as uneven placement of the underwear, ensuring the accuracy and reliability of the pattern layout;

[0033] The starting layout line is selected based on the contour features of the pattern layout area, and then matched with the printing starting line of the pattern layout template; the pattern layout template is proportionally enlarged and the angle features are adjusted to make the two overlap, and finally the complete pattern layout area is determined; this process ensures that the pattern layout template can be accurately applied to the pattern layout area, realizes efficient and accurate pattern layout processing, improves the quality and efficiency of underwear pattern layout, and can meet the needs of refined pattern layout in the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic diagram of the process of the present invention;

[0035] Figure 2 It is a schematic diagram for determining the starting typesetting line of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] See also Figure 1 The present application provides a method for layout of underwear patterns based on machine vision, comprising the following steps:

[0038] Step 1: Place the underwear that needs to be patterned at a designated position, use a high-definition probe to obtain the surface image of the underwear, and perform preliminary processing on the obtained surface image to select a complete image belonging to the underwear. Specifically, after the underwear is placed at the designated position, the corresponding high-definition probe obtains the image of the designated position. Then the obtained image includes not only the relevant image of the corresponding underwear, but also the specific image of the background where the corresponding underwear is located. Therefore, for subsequent feature recognition and specific processing, it is necessary to remove the corresponding background image from the obtained surface image to determine the complete image of the corresponding underwear. The specific sub-steps for selection are:

[0039] S11. Convert the image of the underwear surface obtained by the high-definition probe, converting the original color image (i.e., the underwear surface image) into a grayscale image. Then, confirm the grayscale values associated with different grayscale points in the converted grayscale image, and label different grayscale values as H i , where i represents different grayscale points;

[0040] S12. Compare the grayscale value H i of different grayscale points with the preset value Y1. The specific value of Y1 is determined by the operator according to experience. Label the grayscale points where H i ≤ Y1 as background points, otherwise, do not make any markings;

[0041] Remove the background area associated with the background points from the underwear surface image, and record the image after the removal process as the complete underwear image. Specifically, the grayscale feature differences between the corresponding grayscale image and the background image are relatively obvious. Therefore, the background image and the complete image of the corresponding underwear can be distinguished based on the grayscale features in the corresponding grayscale image, which is convenient for subsequent confirmation of the features of the complete underwear image and completing the overall layout of the relevant pattern matching for the complete underwear image;

[0042] Step 2. Re-convert the determined complete underwear image into a color image, then perform feature contour processing on the converted complete underwear image, identify the pixel values associated with different pixel points in the complete underwear image, and based on the gradient performance of different pixel points, confirm the feature contours existing in the corresponding complete underwear image and perform contour area confirmation. Specifically, different pixel points in the corresponding image have associated pixel values. Based on the change characteristics between the corresponding pixel values of the corresponding pixel points, gradient pixel points can be selected from several groups of pixel points. Then, based on the comprehensive contour performance of the gradient pixel points, the specific area contour can be confirmed, so as to determine the comprehensive contour and internal contour of the corresponding image. Then, specifically confirm the pattern area from the confirmed corresponding contour area. For different types of underwear, based on the different performances of the corresponding contour area, the corresponding pattern area can be locked. The specific sub-steps for confirming the contour area are as follows:

[0043] S21. Label the pixel values associated with different pixel points in the corresponding complete underwear image after conversion as X k , where k represents different pixel points. Then, based on the adjacent pixel points existing around the corresponding pixel points, confirm the pixel values associated with the adjacent pixel points, and use the Sobel algorithm to process the confirmed several groups of pixel values to confirm the horizontal gradient HT k associated with the corresponding pixel point and the vertical gradient ZT k, specifically, when specifically confirming the gradient value of the corresponding pixel point, first determine the horizontal gradient kernel and vertical gradient kernel associated with the corresponding convolution process, and then, based on the specific weight factors associated with the corresponding gradient kernels, assign weights to the pixel values associated with the corresponding pixel points. Then, based on the specific numerical performance states of the pixel points at each different position after being assigned different weights, comprehensively confirm the gradient value associated with the corresponding pixel point, so as to confirm the horizontal gradient and vertical gradient associated with the corresponding pixel point. The associated horizontal gradient kernel is , and the associated vertical gradient kernel is ;

[0044] S22. Based on the confirmed horizontal gradient HT k and vertical gradient ZT k , lock the comprehensive gradient associated with the corresponding pixel point, and use to confirm its comprehensive gradient ZH k . Compare the comprehensive gradient ZH k confirmed for the corresponding pixel point with the preset value Y2, where the specific value of Y2 is determined by the operator according to experience. When ZH k ≤ Y2, no processing is performed. When ZH k > Y2, mark the corresponding pixel point as a gradient pixel point;

[0045] S23. Based on several groups of gradient pixel points confirmed in the complete underwear image, connect adjacent gradient pixel points, and sequentially confirm the characteristic contours existing in the complete underwear image, and then mark the specific areas included in each different characteristic contour as contour areas;

[0046] Specifically, when the underwear is acquiring an image, due to the specific contour features inside and outside it, the internal and external contours of the corresponding underwear can be specifically confirmed. From the several groups of confirmed contours, the different areas of the corresponding underwear can be divided, and then the contour areas are analyzed for features to determine the specific areas that need to be pattern-matched and typeset;

[0047] Step 3. From several groups of contour areas confirmed in the complete underwear image, first determine the center points associated with the corresponding contour areas, and then align the contour areas with the preset characteristic areas and simultaneously perform an overlap analysis to select the pattern-matching and typesetting area in the complete underwear image. The specific method for selection is as follows:

[0048] S31. Based on the different edge contours associated with different contour regions, place the corresponding edge contours in a two-dimensional coordinate system, confirm the two-dimensional coordinates associated with the corresponding contour points, perform mean processing on the several groups of two-dimensional coordinates of the several groups of contour points associated with the corresponding edge contours, confirm the mean coordinates, and mark the points associated with the determined mean coordinates within the corresponding contour region, which is denoted as the center point of this contour region;

[0049] S32. Sequentially confirm the center points within different contour regions, align the contour regions after the center point confirmation with the preset feature region. The preset feature region is the originally planned typesetting region, which is confirmed and input in advance by relevant operators, and the system itself performs feature comparison and coincidence. After alignment, either the contour region rotates according to the center point or the feature region rotates according to the center point (only one of the two needs to rotate). There is also a preset center point within the feature region, and during the rotation process, the overlapping region generated between the contour region and the feature region is recorded in real time. Confirm the ratio of the area of the overlapping region within the feature region, using: ZB p = overlapping region area ÷ total area of the feature region, where p represents different rotation processes. When the contour region or the feature region rotates 360° according to the corresponding center point (i.e., one cycle), stop rotating;

[0050] Based on the different area ratios ZB associated with different rotation processes p , select ZB p max as the characteristic value associated with this contour region;

[0051] S33. Sequentially confirm the characteristic values ZB p max of different contour regions. Mark the contour regions that satisfy: ZB p max ≥ 95% as the pattern alignment typesetting area. Otherwise, do not perform any marking. If there is no contour region in the complete underwear image that satisfies ZB p max ≥ 95%, directly generate an error signal for display, indicating that there is a deviation in the processing process and human intervention is required. Most likely, the corresponding underwear is not placed flat enough, resulting in a large deviation in the area of the corresponding contour region;

[0052] Step Four. Based on the determined pattern alignment typesetting area and the preset pattern template, first determine the relevant characteristics of the overall contour of the pattern alignment typesetting area. Select the starting typesetting line from the pattern alignment typesetting area, and then based on the preset printing starting line in the pattern template, combine the pattern template with the pattern alignment typesetting area, and display the combined complete pattern area. The specific sub-steps for combination are as follows:

[0053] S41. Combine Figure 2, perform a contour confirmation on the determined pattern alignment area, lock the edge contour belonging to this pattern alignment area, and then, from the locked edge contour, confirm the point angle between the corresponding contour points (generally, there are two orientations for the angles corresponding to the three groups of points, an internal angle and an external angle. The point angle determined here is the corresponding angle facing the center point of the corresponding pattern alignment area. Three points determine a group of point angles, and the corresponding contour points are located at the middle points of the three groups of points), and calibrate the different confirmed point angles as Jo, where o represents different corresponding contour points;

[0054] Calibrate the corresponding contour points that satisfy Jo ≤ 150° as characteristic inflection points, otherwise, do not perform any calibration;

[0055] Confirm the length of the part of the edge contour associated between adjacent characteristic inflection points, and calibrate the confirmed length as CL q , where q represents different parts of the edge contour. From the multiple groups of lengths CL q confirmed, select the part of the edge contour associated with CL q max, and denote it as the starting layout line of this pattern alignment area (that is, the longest group of side lines on the edge contour of the corresponding pattern alignment area. The endpoints between the corresponding side lines are all characteristic inflection points. Therefore, when performing layout processing, based on the longest side line for layout processing can achieve a better pattern alignment processing effect);

[0056] S42. Based on a preset pattern template, select a printing starting line from within this pattern template. The printing starting line has been pre - calibrated within the pattern template. Calibrate the length of the printing starting line as YL, and calibrate the length of the starting layout line within the pattern alignment area as QL;

[0057] If YL < QL, then scale up the pattern template proportionally until YL ≥ QL;

[0058] If YL ≥ QL, then do not perform any processing;

[0059] S43. Confirm the center point Z1 of the printing start line and the center point Z of the start typesetting line, make Z1 coincide with Z2, and based on the angular features between the center point Z2 inside the start typesetting line and adjacent points, adjust the angular features between the center point Z1 of the printing start line and adjacent points. Subsequently, taking the center point as the reference, sequentially adjust the angular features of the surrounding points, and adjust step by step in this way until a set of curves inside the printing start line completely coincide with the start typesetting line (corresponding to YL>QL) or the printing start line completely coincides with the start typesetting line (corresponding to YL=QL. In this processing process, it can be equivalent to having a set of straight lines or curves, where the straight line is the printing start line and the curve is the start typesetting line. In order to convert the straight line into a curve, it is necessary to change the angular features of the same points inside the straight line according to the change of the angular features between adjacent points inside the curve, so as to adjust the corresponding straight line into a curve);

[0060] After the printing start line completely coincides with the start typesetting line, make the pattern alignment template coincide with the pattern alignment area, select the coincident area, and use the selected coincident area as the complete pattern alignment area and display it to complete the pattern alignment processing of the specified area of the corresponding underwear;

[0061] Select the start typesetting line based on the contour features of the pattern alignment area, and then match and combine it with the printing start line of the pattern alignment template; make the two coincide by proportionally enlarging the pattern alignment template and adjusting the angular features, and finally determine the complete pattern alignment area; this process ensures that the pattern alignment template can be accurately applied to the pattern alignment area, realizes efficient and accurate pattern alignment processing, improves the quality and efficiency of underwear pattern alignment, and can meet the requirements of fine pattern alignment in the production process.

[0062] Some data in the above formula are numerically calculated after removing their dimensions, and the content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0063] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for layout of underwear patterns based on machine vision, characterized in that: The following steps are involved: Step 1: Place the underwear to be patterned at a designated position, use a high-definition probe to obtain a surface image of the underwear, and perform preliminary processing on the obtained surface image to select a complete image of the underwear; Step 2: reconvert the determined complete underwear image into a color image, and then perform feature contour processing on the converted complete underwear image to identify the pixel values ​​associated with different pixel points in the complete underwear image, and then confirm the feature contour existing in the corresponding complete underwear image based on the gradient performance of different pixel points, and confirm the contour area; Step 3: From among the several groups of contour areas confirmed in the complete underwear image, the center points associated with the corresponding contour areas are preferentially determined, and then the contour areas are aligned with the preset feature areas and coincidence analysis is performed simultaneously, and the pattern layout area is selected in the complete underwear image; Step 4. Based on the determined pattern layout area and the preset pattern layout template, prioritize the relevant features of the overall outline of the pattern layout area, select the starting layout line from the pattern layout area, and then combine the pattern template with the pattern layout area based on the preset printing starting line in the pattern template: calibrate the lengths of the printing starting line and the starting layout line, determine the center point, adjust the angle feature between the center point of the printing starting line and the adjacent points until it coincides with the starting layout line, so that the pattern template coincides with the pattern layout area, and display the combined complete pattern area.

2. The method for underwear pattern layout based on machine vision according to claim 1, characterized in that: In the step 1, the specific sub-steps of selecting the complete image belonging to the underwear are: S11, converting the underwear surface image obtained by the high-definition probe to convert the original color image into a grayscale image, confirming the grayscale values ​​associated with different grayscale points in the converted grayscale image, and calibrating the different grayscale values ​​as H i , where i represents different grayscale points; S12, the gray value H of different gray points i Compare with the preset value Y1 and set H i Grayscale points ≤Y1 are marked as background points, otherwise, no marking is performed; The background area associated with the background points is removed from the underwear surface image, and the image after the removal is recorded as the complete underwear image.

3. The method for underwear pattern layout based on machine vision according to claim 1, characterized in that: In step 2, the specific sub-steps of confirming the contour area are: S21, calibrating the pixel values ​​associated with different pixel points in the complete underwear image after the conversion processing as X k , where k represents different pixels, and then based on the adjacent pixels around the corresponding pixel, the pixel values ​​associated with the adjacent pixels are confirmed, and then the Sobel algorithm is used to process the confirmed groups of pixel values ​​to confirm the lateral gradient HT associated with the corresponding pixel. k and the longitudinal gradient ZT k ; S22, based on the confirmed transverse gradient HT k and the longitudinal gradient ZT k , lock the comprehensive gradient associated with the corresponding pixel point, using Confirm its comprehensive gradient ZH k , the comprehensive gradient ZH confirmed by the corresponding pixel point k Compare with the preset value Y2, when ZH k >Y2, the corresponding pixel point is marked as a gradient pixel point; S23, based on the confirmed groups of gradient pixel points in the complete underwear image, adjacent gradient pixel points are connected, characteristic contours existing in the complete underwear image are confirmed in sequence, and then specific areas contained in each different characteristic contour are marked as contour areas.

4. The method for underwear pattern layout based on machine vision according to claim 3, characterized in that: In step S22, when ZH k When ≤Y2, no processing is performed.

5. The method for underwear pattern layout based on machine vision according to claim 1, characterized in that: In the step 3, the specific method of selecting the pattern layout area in the complete underwear image is: S31, based on different edge contours associated with different contour areas, placing the corresponding edge contours in a two-dimensional coordinate system, and confirming the two-dimensional coordinates associated with the corresponding contour points, performing mean processing on several groups of two-dimensional coordinates of several groups of contour points associated with the corresponding edge contours, confirming the mean coordinates, and marking the point associated with the determined mean coordinates in the corresponding contour area, and recording it as the center point of this contour area; S32, confirm the center points inside different contour areas in turn, align the center points of the contour area after the center point confirmation with the preset feature area, where the preset feature area is the originally planned layout area, and after alignment, rotate the contour area according to the center point or the feature area according to the center point, where the center point is preset inside the feature area, and record the overlapped area between the contour area and the feature area in real time during the rotation process, confirm the area ratio of the overlapped area in the feature area, and use: ZB p = overlapping area ÷ total area of ​​feature area, where p represents different rotation processes. When the contour area or feature area rotates 360° based on the corresponding center point, it stops rotating; Different area ratio values ​​ZB associated with different rotation processes p , select ZB p max is the eigenvalue associated with this contour area; S33, the characteristic value ZB associated with different contour areas p max is confirmed in turn, which will satisfy: ZB p The contour area with max ≥ 95% is calibrated for the layout area, otherwise no calibration is performed.

6. The method for underwear pattern layout based on machine vision according to claim 5, characterized in that: In step S33, if ZB does not exist in the complete underwear image, p For contour areas with max≥95%, an error signal is directly generated for display.

7. The method for underwear pattern layout based on machine vision according to claim 1, characterized in that: In the step 4, the specific method of selecting the starting layout line from the pattern layout area is: S41, confirming the contour of the determined pattern layout area, locking the edge contour belonging to the pattern layout area, and then confirming the point angles between corresponding contour points from the locked edge contour, and calibrating the confirmed different point angles as Jo, where o represents different corresponding contour points; The corresponding contour points that satisfy Jo≤150° are calibrated as characteristic inflection points, otherwise no calibration is performed; Confirm the length of the partial edge contour associated between adjacent feature inflection points, and mark the confirmed length as CL q , where q represents different partial edge profiles, from the multiple sets of lengths CL identified q In the q The partial edge contour associated with max is recorded as the starting layout line of this layout area.

8. The method for underwear pattern layout based on machine vision according to claim 7, characterized in that: In step 4, the specific method of determining the complete area of ​​the flower is: S42, based on a preset pattern alignment template, selecting a printing start line in the pattern alignment template, calibrating the length of the printing start line as YL, and calibrating the length of the start layout line in the pattern alignment area as QL; If YL<QL, the pattern template is enlarged proportionally until the length of the printing starting line YL≥QL; If YL ≥ QL, no treatment is performed; S43, confirming the center point Z1 of the printing starting line and the center point Z2 of the starting layout line, making Z1 coincide with Z2, and adjusting the angle feature between the center point Z1 of the printing starting line and the adjacent points based on the angle feature between the center point Z2 and the adjacent points in the starting layout line, and then adjusting the angle feature of the surrounding points in turn based on the center point, and gradually adjusting in this way until a group of curves in the printing starting line completely coincide with the starting layout line or the printing starting line completely coincides with the starting layout line, and stopping; After the printing start line and the start layout line are completely overlapped, the pattern template and the pattern layout area are overlapped, and the overlap area is selected, and the selected overlap area is used as the complete pattern area and displayed.

Images